A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents.pdf

Running Head: META-ANALYSIS OF RISK FACTORS FOR PTSD
A Meta-Analysis of Risk Factors for Posttraumatic Stress Disorder in Children and
Adolescents
David Trickeya
, Andy P. Siddawayb
, Richard Meiser-Stedmanc
, Lucy Serpelld,e
, Andy
P. Fieldf
a
Child Bereavement & Trauma Service (CHUMS), Cambridgeshire Community
Services NHS, UK
b
School of Psychology, University of Hertfordshire, UK
c
MRC Cognition and Brain Sciences Unit, Cambridge, UK
d
Research Department of Clinical, Educational & Health Psychology, University
College London, UK
e
North East London NHS Foundation Trust
f
School of Psychology, University of Sussex, UK
Correspondence to: David Trickey, Child Bereavement & Trauma Service (CHUMS
Social Enterprise CIC), Sundon Park Health Centre, Tenth Avenue, Luton, LU3 3EP, UK.
mail@davidtrickey.com, +44 7734432222, fax +44 1372353367

META-ANALYSIS OF RISK FACTORS FOR PTSD 2
Abstract
Post traumatic stress disorder (PTSD) is a complex and chronic disorder that causes
substantial distress and interferes with social and educational functioning. Consequently,
identifying the risk factors that make a child more likely to experience traumatic distress is of
academic, clinical and social importance. This meta-analysis estimated the population effect
sizes of 25 potential risk factors for PTSD in children and adolescents aged 6-18 years across
64 studies (N = 32,238). Medium to large effect sizes were shown for many factors relating to
subjective experience of the event and post-trauma variables (low social support, peri-trauma
fear, perceived life threat, social withdrawal, comorbid psychological problem, poor family
functioning, distraction, PTSD at time 1, and thought suppression); whereas pre-trauma
variables and more objective measures of the assumed severity of the event generated small
to medium effect sizes. This indicates that subjective peri-trauma factors and post-event
factors are likely to have a major role in determining whether a child develops PTSD
following exposure to a traumatic event. Such factors could potentially be assessed following
a potentially traumatic event in order to screen for those most vulnerable to developing PTSD
and target treatment efforts accordingly. The findings support the cognitive model of PTSD
as a way of understanding its development and guiding interventions to reduce symptoms.
Keywords: Posttraumatic Stress Disorder, risk factor, Children, Adolescent, predictor

A Meta-Analysis of Risk Factors for Posttraumatic Stress Disorder in Children and
Adolescents
Psychological reactions of children and adolescents to potentially traumatic events
have been studied for more than 20 years. Although other reactions such as depression,
anxiety and behaviour problems do occur, the most commonly studied reaction is
Posttraumatic Stress Disorder (PTSD). PTSD can be a complex and chronic disorder that
commonly co-occurs with other disorders including other anxiety disorders, mood disorders
and substance use disorders (Brewin, Dalgleish & Joseph, 1996). Symptoms of PTSD can
cause substantial distress and interfere with social and educational functioning (National
Institute for Health & Clinical Excellence, 2005). A meta-analysis of rates of PTSD in
children and adolescents exposed to trauma, which combined the results of 34 studies,
including a total of 2697 individuals found that overall, 36% were diagnosed with PTSD
(Fletcher, 1996). However rates vary widely across studies from 0% to 100% (Dalgleish,
Meiser-Stedman & Smith, 2005), indicating that exposure to a potentially traumatic event is
necessary, but not sufficient for a young person to develop PTSD. It is therefore increasingly
accepted that factors other than the potentially traumatic event play a role in explaining PTSD
in both adults and children (e.g. Brewin, Andrews & Valentine, 2000; Foy, Madvig, Pynoos
& Camilleri, 1996).
Rationale for the Present Study
Knowing reliably which factors make a child more likely to develop traumatic distress
(whether that is measured by a diagnosis of PTSD, or by severity of PTSD symptoms) is of
both academic and clinical interest. Firstly if clinicians know which children and adolescents
are most likely to be adversely affected following exposure to events, then those children and
adolescents can be most closely monitored with a view to providing treatment as necessary
before difficulties become chronic (Pine & Cohen, 2002). Secondly, robust knowledge of risk

factors may help our understanding of PTSD and its causes, by confirming or disconfirming
current models. In turn, a better understanding of PTSD may assist in assessment, prevention
and intervention leading to better outcomes for children and young people exposed to
potentially traumatic events.
Thirdly, such knowledge will also highlight areas for further research.
The extent to which knowledge can be robust relates not just to obtaining precise
estimates of population effect sizes for risk factors, but also estimating the bias in these
estimates. One likely source of bias is publication bias (the ‘file drawer problem’): studies
that yield non-significant risk factors may be less likely to be submitted for publication;
where multiple risk factors are investigated authors may choose to report only the statistically
significant one; and editors and reviewers may be unwilling to publish non-significant
findings or ask for non-significant effects to be trimmed to save space. Therefore, it is
important to estimate the likely effects of publication bias on the conclusions we draw about
risk factors.
Existing Research
In trying to examine how different factors may be associated with the development of post-
traumatic distress, we have categorised the wide range of different factors into the following
categories: demographic factors, pre-trauma factors, objective trauma characteristics,
subjective trauma characteristics, post-trauma individual factors, post-trauma psychological
environment. The research relating to each of these categories is outlined in more detail
below.
Demographic factors.
Demographic characteristics have been examined in many studies. While data
concerning these factors may in themselves not lend support to any particular theoretical
account of PTSD, they are easy to assess and thus may aid efforts to identify youth at high

risk of developing PTSD following exposure to a potentially traumatic event. Age has
received particular attention and although results are mixed (Foy et al., 1996), a number of
mechanisms have been proposed for how age may alter a young person’s vulnerability to
developing PTSD (e.g. Salmon & Bryant, 2002). Gender has also been considered to play a
role in the development of symptomatic distress, and it has been suggested that this may be
due to female survivors exhibiting more extreme acute reactions which may lead to increased
risk of later PTSD (Pine & Cohen, 2002).
Pre-trauma factors
The question of how pre-trauma factors, such as previous traumas, may increase a
young person’s vulnerability or resilience following a potentially traumatic event has been
explored in a number of studies. Viewing a child’s response to a traumatic stressor within the
context of their pre-trauma life experiences and mental health is consistent with a
developmental psychopathological perspective, and may identify important risk factors for
the development of chronic PTSD. Foy and colleagues have suggested several ways in which
various pre-trauma factors may play a role in the development of PTSD. Some factors may
interact with the effects of the traumatic experience to heighten reactivity and thereby
increase an individual’s vulnerability, whilst others may increase distress independent of the
effect of the traumatic experience. They go on to suggest that other factors may act both by
interacting with the impact of the event and causing independent distress (Foy et al., 1996).
Objective trauma characteristics
Several commentators have noted that objective characteristics such as severity of
event, or level of exposure, consistently predict subsequent PTSD symptoms (Pine & Cohen,
2002, Foy et al., 1996). Quantifying the strength of association between objective trauma
characteristics and PTSD allows for an examination of the relative importance of such trauma

characteristics compared to other factors, as well as the identification of further ways of
identifying youth at high risk.
Subjective trauma characteristics
Within the current DSM-IV criteria for PTSD, subjective characteristics of an event
(the experience of intense fear, helplessness or horror) – are an essential requirement for
conferring this diagnosis. Some commentators have questioned whether this criterion is
appropriate for young children given the difficulties inherent in assessing peri-trauma
reactions in this age group (Scheeringa et al., 2003). However, subjective factors, such as the
perception of threat and peri-traumatic affect, are highly likely to be linked to the onset of
PTSD (Perrin et al., 2000). Indeed, perceived life threat and fear have already been included
in screening measures for trauma-exposed youth (Winston, Kassam-Adams, Garcia-España,
Ittenbach, & Cnaan, 2003).
Post-trauma individual factors
Comorbid psychological difficulties, acute post-traumatic stress symptoms, and
coping styles are among the factors intrinsic to the child that have been considered as risk
factors for PTSD. Other cognitive (Dalgleish et al., 2005; Meiser-Stedman, 2002; Salmon &
Bryant, 2002) and biological mechanisms (De Bellis et al., 1999) have also been suggested as
playing an important role in the aetiology of PTSD. However, it remains to be seen whether
there is any consistent pattern in the literature concerning such factors.
Post-trauma psychological environment
Pine and Cohen’s review found that in addition to trauma exposure, levels of social
support were consistently associated with mental health outcomes for children exposed to a
potentially traumatic event (Pine & Cohen, 2002). Furthermore, Pynoos and colleagues
suggested that child-extrinsic factors such as parental and family functioning may be

associated with a child’s vulnerability by moderating the impact of the event, the impact of
proximal reminders and the impact of secondary stresses (Pynoos et al., 1999).
Existing reviews of research.
There are a number of reviews of the literature that have attempted to increase our
understanding of the development of child PTSD (Foy et al., 1996; Pine & Cohen, 2002).
Pine and Cohen’s review identified two factors consistently related to adverse outcome:
trauma exposure and disrupted social support. Foy also argued that parental distress might act
as a powerful mediating factor in the development of the child’s symptoms. Whilst such
reviews are helpful, a more systematic review and meta-analysis of factors considered by
research would add significantly to our understanding by enabling a closer examination of the
consistency of the relationship between a particular factor and outcome across different
studies. It would also enable results from different studies to be combined to produce a more
accurate estimate of the population effect size for different risk factors.
Individual studies may report very different strengths of associations between a given
risk factor and PTSD. Meta-analyses are able to explore this variation between studies, by
examining how sample and study characteristics act as moderators to the association.
Other meta-analyses of risk factors for PTSD.
There are two meta-analyses of studies examining risk factors for PTSD in adults
(Brewin et al., 2000 and Ozer, Best, Lipsey & Weiss, 2003). Brewin and colleagues
conducted meta-analyses of 14 risk factors and analysed moderating effects of various sample
and study characteristics. The effect size of all the risk factors was modest; but factors
operating during or after the trauma, such as trauma severity, lack of social support, and
additional life stress, had larger effect sizes (r = .23 to .40) than pre-trauma factors (r = .05 to
.19). Their analysis of moderators revealed that some risk factors (such as gender, age at
trauma and race) predicted PTSD only in some populations whilst some predicted PTSD

more consistently (such as education, previous trauma, and general childhood adversity).
Results varied depending on the population studied (e.g. civilian or military) and the methods
used; and some factors (such as psychiatric history, reported childhood abuse, and family
psychiatric history) predicted PTSD more consistently. Ozer and colleagues conducted a
meta-analysis of data concerning seven separate risk factors, all of which produced
significant but modest effect sizes. Peri-traumatic psychological processes rather than prior
characteristics were found to be the strongest predictors of PTSD symptoms or diagnostic
status.
Cox and colleagues reviewed the literature concerning risk factors following
accidental trauma in children and adolescents and found that eight factors had been measured
in studies that met their inclusion criteria for their meta-analysis (Cox, Kenardy & Hendrikz,
2008). Whilst all risk factors were statistically significant, only threat to life (r = .38) and pre-
trauma psychopathology (r = .29) were considered to be strong predictors of PTSD. Female
gender (r =.18) and pre-trauma psychopathology (r = .22) were weak predictors, whilst
younger age (r = -.04), exposure to prior trauma (r = .08), injury severity (r = .09), and
involvement of family or a friend in the accident (r = .09) were even weaker predictors. Most
risk factors were significantly varied across studies with only pre-trauma psychopathology
and threat to life showing consistent effect sizes across studies. Restricting their analysis to
studies of accidental trauma may have limited the generalisability of results. The present
study therefore included research that investigated a broader range of events, including
natural events, and intentional violence.
Brewin et al. (2000) and Cox et al. (2008) both employed Rosenthal’s fixed-effects
model of meta-analysis (Rosenthal 1991). However, random-effect meta-analyses are
generally more appropriate for several reasons (see Field, 2001; 2003; 2005; Hunter &
Schmidt, 2000). First, real-world data are likely to have variable population parameters

(Field, 2003; Hunter & Schmidt, 1990, 2000; National Research Council, 1992; Osburn &
Callender, 1992). Second, Hedges and Vevea (1998) have suggested that fixed-effect models
are appropriate only when making inferences about the studies included in the meta-analysis
and generally researchers want to generalize their conclusions beyond the studies within the
analysis. Finally, applying fixed effects meta-analytic models to random-effects data can
have undesirable effects on the outcomes of analyses but the reverse is not true (Field, 2003;
Hunter & Schmidt, 2000). Using Monte Carlo analysis Field (2003) demonstrated that fixed-
effects model are unreliable (they inflate estimates of population effect-sizes and yield
corresponding confidence intervals that are too narrow) when there is substantial or moderate
heterogeneity: the chance of a Type I error rose from 5% to between 43 and 80%. For these
reasons, we considered it important to employ a random-effects model.
The present study adds to the meta-analyses described above because it brings
together all of the existing child and adolescent research of sufficient quality that investigates
risk factors for PTSD following a variety of potentially traumatic events using random-effects
meta-analytic techniques.
Method
Selection of Studies for the Meta-analysis
English-language articles published in peer-reviewed journals between 1980 (the first
DSM definition of PTSD) and May 2009 were considered for inclusion. Various
psychological and medical literature databases were searched, including the PILOTS database
managed by the National Center for PTSD, Medline, PsychInfo, Embase and Web of
Science. In addition, each issue of the Journal of Traumatic Stress was examined for relevant
research. Secondary sources such as review articles, book chapters and the reference sections
of selected articles were examined. Search terms for the literature databases included
combinations of the following: PTSD, posttraumatic stress disorder or post-traumatic stress

disorder; child or children or adolescent(s); and, risk, predictor, prediction or predisposition.
A predictor or risk factor was operationally defined as any variable examined as a potential
contributor to variability in PTSD symptom level or diagnostic status. This literature search
yielded a preliminary database of 1210 published articles, which were then reviewed for
inclusion in the meta-analysis using various inclusion and exclusion criteria.
Inclusion and Exclusion Criteria
There is some debate in the child PTSD literature as to whether PTSD manifests
differently in children of different ages and developmental levels, and therefore how best to
assess PTSD in children of different ages (e.g. Dehon & Scheeringa, 2006; Fletcher, 1996;
Meiser-Stedman, Smith, Glucksman, Yule & Dalgleish, 2008; Scheeringa, Zeanah, Myers &
Putnam, 2003). With the exception of assessments specifically for pre-school children (e.g.
Scheeringa, 2003) most standardised assessments of PTSD in children have been validated on
groups of children of school age and above. Thus, the age range for samples to be included in
the present meta-analysis was set at 6-18 years as long as the measure of PTSD was
appropriate for the age group of the sample. A more conservative approach would have been
to set the lower age range at 8 years, however a random effect moderator analysis of effect-
sizes across all risk factors demonstrates that there was no significant difference in overall
effect-size between studies that breached the 8-18 years age range, and those that did not,
(χ2
(1) = .079, p = .778)
Bal et al. (2003) breached the upper age limit by just one year with their age range of
11 – 19. We decided to include this study because the vast majority of the sample was within
the age range, and those that were not, were only just above it.
To be included in the analysis, studies had to have used measures of child PTSD that
considered all three PTSD symptom clusters (intrusion, avoidance/numbing and
hyperarousal), and demonstrated adequate reliability and validity as shown by publication of

their psychometric properties in a peer-reviewed journal or by having a strong pedigree (e.g.
minimal but necessary changes to a measure with proven reliability and validity such as the
UCLA PTSD Reaction Index which has been modified in order to take into account changes
in diagnostic criteria, Steinberg, Brymer, Decker & Pynoos, 2004).
Articles were excluded on any of the following grounds:
(a) The study measured only acute trauma response (e.g. Acute Stress Disorder
or PTSD measured before 1 month post-trauma) rather than PTSD, which
according to DSM-IV-TR, can be diagnosed only after 1 month (American
Psychiatric Association, 2000)
(b) The study used a categorical measure of PTSD and included in the same
comparison group individuals meeting full diagnostic criteria for PTSD
together with individuals with less severe posttraumatic symptoms or
partial PTSD (e.g. ‘subsyndromal PTSD’), and contrasted them with a
group exposed to the same event but without PTSD;
(c) The study sample consisted entirely of individuals who fulfilled the full
diagnostic criteria for PTSD and therefore were unsuitable for ascertaining
the frequencies of these phenomena
(d) The study sample was selected on the basis that participants were
experiencing a specific comorbid psychiatric disorder (e.g. depression,
ADHD, substance abuse, learning difficulties, offenders), which would
limit the generalizability of results
(e) The study did not specifically assess DSM-defined PTSD symptoms (e.g.
studies that reported only general symptoms)
(f) Studies with insufficient data to calculate univariate effect sizes, and where
such data could not be obtained from the study author

(g) The article was a review that did not present new data or only presented
qualitative analyses
(h) The primary aim of the study was to investigate the efficacy of treatment;
and
(i) Single case studies.
Finally any study that examined a risk factor that was not also examined by at least
one other study was excluded from further examination. All eligible studies were carefully
reviewed by two authors (LS & AS) to ensure decision-rule consistency, with 100%
agreement.
Coding of Studies
Following this procedure, 62 studies with a total sample size of 32,238 participants
yielded 453 effect sizes that were included in the meta-analysis. These studies are identified
with an asterisk in the reference section. Table 1 provides a full list of the data extracted from
each study for each risk factor and Table 2 provides a detailed list of characteristics of the
studies used in the meta-analysis. Effect-sizes for both current and past PTSD symptoms, as
well as longitudinal studies that presented multiple assessments of symptoms over time, were
included. The simple mean was computed for studies contributing multiple effect-sizes for
the same risk factor. We did not include multiple effect-sizes from the same study because
the results could be biased by the studies contributing numerous effect sizes. Only single-df
comparisons were appropriate for selection in the meta-analysis (Rosenthal, 1991). Different
articles reporting analyses from the same data set were included if the studies provided effect-
size estimates for different risk factors. On nine occasions, the same data were reported in
more than one publication. When this occurred, estimates from the largest sample or from the
most comprehensive article were used.

For obvious reasons, meta-analysis was not conducted on any risk factor for which
only a single effect size was available (e.g. cortisol levels, handedness, dissociation and grief
reactions risk factors). Where continuous (symptom severity) and categorical (diagnosis)
measures of a given risk factor were both available within a study, we used the effect-size for
PTSD symptom severity because of the statistical advantages of continuously measured
variables in predictive research. Whenever the same risk factors had been assessed within a
study using more than one measure (e.g., several separate aspects of previous psychiatric
history might have been assessed), effect sizes were averaged across these measures. An
effect size of zero was assigned for the few studies (k = 2, 0.4% of all effect sizes) that
reported non-significant findings and did not provide an effect-size. Although this
conservative strategy generally underestimates the true magnitude of effect sizes (Durlak &
Lipsey, 1991; Rosenthal, 1995), this approach is preferable to excluding non-significant
results from the meta-analysis, which would result in an overestimation of combined effect
sizes (Rosenthal, 1995).
Computation and Analysis of Effect Sizes
Meta-analyses were conducted using SPSS 18 and R 2.10 (R Core Development
Team, 2010) using Field and Gillett’s (2010) syntax. A separate meta-analysis was carried
out for each risk factor. In the present study, Pearson’s correlation coefficient, r, was chosen
as the effect size for a number of reasons. First, r is a common metric for which the greatest
number of effect sizes could be reported or converted; second, r is easily computed from chi-
square, t, F, and d (see Hunter & Schmidt, 2004); and third, r is readily interpretable in terms
of practical importance (Field, 2001; Rosenthal, 1991; Rosenthal & DiMatteo, 2001).
For categorical data, correlation coefficients were computed such that a positive
coefficient reflected a higher mean in the PTSD group than the control group, and a negative
coefficient represented a lower mean in the PTSD group than the control group. For

continuous data, correlation coefficients were computed such that a positive coefficient
reflected more severe PTSD symptoms, and a negative coefficient reflected less severe PTSD
symptoms. Higher values of r indicate a stronger positive association with PTSD. Table 1
shows the effect sizes included in the meta-analysis for each risk factor.
Insert Table 1
Method of Meta-Analysis
For reasons discussed earlier, a random-effects meta-analysis was used. Hedges’
(Hedges and Olkin, 1985; Hedges & Vevea, 1998) method was applied using Fisher-
transformed correlation coefficients with results reported after the back transformation to the
Pearson product–moment correlation coefficient (see Field, 2005; Overton, 1998). Using this
method, each effect size is weighted by a value reflecting both the within study variance
(1/n−3 for correlation coefficients in which n is the sample size) and the between study
variance (τ2
) (see Field & Gillett, 2010, for a guide to using Hedges and Vevea's method).
Moderator analyses were conducted using a random-effects general linear model in
which each z-transformed effect size can be predicted from the transformed moderator effect
(represented by regression coefficient, β). The moderator effect, β, is estimated using
generalised least squared (GLS). In both the main analysis and moderator analyses, between
study variance was estimated non iteratively (e.g. Dersimonian & Laird, 1986). For a
technical overview of the GLS moderator analysis that we employed see Overton (1998) or
Field and Gillett (2010).
Results
Characteristics of studies
The method described above generated 25 risk factors that were explored by two or
more studies, which could then be entered into the meta-analysis. Sample sizes from

individual studies ranged from 20 to 5,687. Characteristics of the studies included in the
meta-analysis (trauma type, sample size, measure of PTSD, whether the measure was an
interview or questionnaire, age range, mean age, percentage of sample that were female, and
location of study) are contained in Table 2.
Insert Table 2
Risk factor estimates
The main results of the meta-analysis are for each risk factor are contained in Table 3.
For each risk factor this table shows the number of studies (k), the estimate of between-study
variability (τ2
), test of significance of between study variability (χ2
), estimate of the
population effect size (𝜌
̂), 95% confidence intervals, test of the population effect size (z) and
estimate of the population effect size under severe two-tailed publication bias (pb).
Insert Table 3
The estimate of the population effect size, the confidence intervals and the range of
effect sizes from individual studies are shown diagrammatically in figure 1.
Insert Figure 1
Two risk factors yielded population effect size estimates close to zero: race (Black
and minority ethnic; BME) and younger age. However, the effect of race, whilst small, was
nevertheless statistically significant unlike that of younger age, which was not.
Thirteen risk factors yielded small to medium population effect size estimates (i.e.
with an absolute value between .1 and .3, explaining 1-9% of the variance in PTSD

symptoms). It is worth noting that the effect of media exposure was not statistically
significant, but this is likely to reflect the small number of studies on which the effect size
estimate were based (k = 3). The remaining 10 risk factors yielded medium to large effect
sizes (i.e. greater than .3, explaining more than 9% of the variance in PTSD symptoms).
There are some noteworthy caveats to these results. First, most of these effects are
based on very small numbers of studies. For example, the two largest effect size estimates
(thought suppression and PTSD at time 1) are based on only 2 and 4 studies respectively. As
such, these effect sizes could reflect idiosyncrasies in the studies, or may simply not be
generalizable. The only risk factor to yield a relatively large effect size estimate that was
based on a large number of studies was post-trauma psychological problem (k = 25). Second,
three risk factors showed significant heterogeneity across effect sizes from individual studies
(trauma severity, blame others and post-trauma psychological problem). However this is in
contrast to other meta-analysis of risk factors for PTSD (e.g. Brewin et al., 2000 and Cox et
al., 2008) in which a majority of factors had varied results from different studies. The
heterogeneity of the effect sizes for trauma severity and post-trauma psychological problem
can in part be explained by the relatively large number of studies, (k = 41and 25
respectively), the estimate of variability, τ2
, itself was not large in either of these cases and so
the significance could reflect the relatively large power of the test.
Moderator Analysis
Four risk factors (trauma severity, comorbid psychological problems, younger age,
and female gender) had sufficient numbers (k ≥ 18 in the smallest category) of studies to
enable an analysis of variables that might moderate the effect sizes. Five potential moderator
variables were identified:

(1) Whether PTSD was measured by interview or questionnaire (which in Brewin et
al., 2000, moderated results for female gender, younger age at trauma, trauma severity and
previous trauma).
(2) Whether the trauma was a group trauma (i.e., the same event affecting very many
people, such as an earthquake) or an individual trauma (an event happening to one individual,
such as an assault). It is possible that social support following a group trauma manifests in
different ways to social support following individual events, and given the importance of
social support in Brewin’s meta-analysis of adult risk factors, this was considered to be of
interest as a possible moderating factor.
(3) Whether the trauma was intended (e.g., war, terrorism, abuse, shooting) or not. In
Fletcher’s meta-analysis (1996) rates of PTSD were found to be different depending on
whether the event was a natural event or “human-caused”, which may indicate that different
meaning is attributed to different types of events, which have subsequent effects on outcome.
Categorising events as either intended or not was considered to be a more discriminating
categorisation than natural or “human-caused”, as the fact that the harm was caused on
purpose may be even more meaningful for the victims, than simply the fact that it was human
caused.
(4) Whether the measure of PTSD was categorical or continuous. This was a
significant moderator of six factors in Brewin et al. (2000).
(5) The mean age of the sample (which was considered worthy of further
investigation given the discussion of age and PTSD in the introduction).
For comorbid psychological problem we used an additional moderator variable that
specified the type of psychological problem in three categories: anxiety, depression and
‘other’.

Table 4 shows the results of all moderator analyses, including betas for the moderator,
95% CIs, and their significance. These analyses will be discussed for each risk factor in turn.
Tables 5-8 show the individual meta-analyses for these risk factors by moderating group
(except for mean age, which was a continuous predictor and needed no sub-analysis). In all
analyses, intended trauma had three groups (intended, unintended and mixed) and so was
entered as two dummy variables (intended vs. unintended, intended vs. mixed), in this case
the overall moderation effect is reported in the text (as Q), and the individual contributions of
the dummy variables are in Table 4.
Insert Table 4
Trauma severity.
For Trauma severity, there were no significant moderators of the population effect
size estimate i.e. the effect size for each group was not significantly different from their
comparison group (Table 4). Although somewhat higher effect sizes were found for
questionnaires (compared to interview), mixed trauma (compared to intended and
unintended), and group trauma (compared to individual). The effect size for each group
individually was significant (Table 5). The differences between groups may not have been
statistically significant, however the fact that one is bigger than the other is still of interest.
Insert Table 5
Comorbid psychological problems.
For the additional moderator of the type of psychological problem two dummy
variables representing anxiety compared to ‘other’, and depression compared to ‘other’ were
included. Overall, there was a significant moderation effect Q(2) = 7.15, p = .028. Table 4
shows that effect sizes for depression differed significantly to other, p = .008. For anxiety,

effects were higher than for other, although this did not quite reach statistical significance, p
= .056. Table 6 shows that the population effect size estimate for ‘other’ was significantly
lower than both anxiety and depression; however, population effect size estimates were
significantly different from zero in all three groups (anxiety, depression and other).
The only other significant moderator was whether the trauma was intended or not.
Intended trauma gave rise to significantly higher effect sizes than unintended trauma,
although both population effect size estimates were significantly greater than zero (Table 6).
Insert Table 6
Younger age.
For the younger age risk factor, intentionality of trauma was a significant moderator
overall, Q(2) = 10.70, p < .005. The dummy variables (Table 4) and sub-group analyses
(Table 7) showed that effect sizes for intended trauma were significantly lower (and in the
opposite direction) than both unintended trauma and mixed trauma. Mixed trauma was the
only subgroup for which the population effect size estimate differed significantly from zero,
however small numbers of studies indicate caution when drawing conclusions from these
results. The only other significant moderator was group vs. individual trauma (Table 4).
Effect sizes were significantly larger for group trauma compared to individual trauma (Table
7), and the population effect size differed significantly from zero for group trauma, but not
for individual trauma.
Insert Table 7
Female gender.
For the female gender risk factor, intentionality of trauma was a significant moderator
overall, Q(2) = 7.30, p = .026. The dummy variables (Table 4) and sub-group analyses (Table
8) showed that effect sizes for intended trauma were significantly lower than both unintended
trauma and mixed trauma. Population effect size estimates differed significantly from zero in

all three groups (Table 8). The only other significant moderator was mean age: the positive
beta (Table 4) indicates that effect sizes increased significantly as a function of mean age. (In
other words, the degree to which being female significantly increases risk increases with
age.)
Insert Table 8
Publication Bias
‘Publication bias’ (or the so-called ‘file-drawer problem’) describes the relationship
between the decision to publish a paper and the results of the paper (Begg, 1994). Studies
with non-significant findings are less likely to be published than studies with significant
findings, which could result in a positive bias within the child PTSD literature. In addition,
authors of studies that measure multiple risk factors may choose to report only those that
were significant.
Although Rosenthal’s fail-safe N is commonly used to quantify publication bias
(Rosenthal, 1979), it is problematic because it emphasizes significance testing the population
effect size rather than estimating bias in the population effect size itself. We performed a
sensitivity analysis, which ‘corrects’ the population effect size estimate for publication bias
based on a weight function that reflects the nature of the bias. We applied Vevea and Woods’
(2005) methods because they can be applied to relatively small samples of studies. In Table 3
we report pb, which is an estimate of the population effect size under severe two-tailed
publication bias. (We corrected for other models too, but the results were consistent with the
reported model.) In Table 3, if pb and 𝜌
̂, are similar, then publication bias has had little
effect. Table 3 shows that although the population effect size estimates are reduced when
corrected for publication bias, our basic conclusions remain unchanged.

Discussion
This meta-analysis involved investigating 25 risk factors for PTSD in children and
adolescents aged 6-18 years, across 64 studies published between 1980 and May 2009. The
combined sample size for the studies considered was very large (32,238). A small effect size
was observed for race and younger age as risk factors for PTSD; a small to medium-sized
effect was observed for female gender, low intelligence, low SES, pre- and post-trauma life
events, pre-trauma psychological problems in the individual and parent, pre-trauma low self-
esteeem, post-trauma parental psychological problems, bereavement, time post-trauma (an
inverse relationship), trauma severity, and exposure to the event by media; while a large
effect was observed for low social support, peri-trauma fear, perceived life threat, social
withdrawal, comorbid psychological problem, poor family functioning, distraction, PTSD at
time 1, and thought suppression. These findings are consistent with the earlier reviews of Foy
and colleagues (1996) and Pine and Cohen (2002), but go further in quantifying the relative
importance of these risk factors. It should be noted that in discussing and comparing the
effect sizes of different factors, there is no implication that the difference is necessarily
statistically significant.
However, only 6 out of the 25 variables examined were investigated in 10 or more
studies. Despite the increase in the number of studies into child PTSD over the past 25 years,
only a limited number of variables have been routinely investigated. In contrast, in Brewin
and colleagues’ meta-analysis examining factors related to PTSD in adult samples, 11 of the
14 variables under consideration had been examined in 10 or more studies (Brewin et al.,
2000). The present meta-analysis therefore highlights the need for the further investigation of
some very rudimentary potential risk factors (e.g. low intelligence, race, perceived life threat,
low social support). Furthermore, it highlights some areas for future research about which we
currently know relatively little but, based on preliminary evidence, may turn out to be very

important. Examples of such factors are thought suppression, media exposure and social
withdrawal.
In order to consider the impact on our results of publication bias, we estimated
population effect sizes under severe two-tailed publication bias. This suggested that the above
pattern of results was not substantially affected by publication bias. While publication bias
may not, therefore, have been a major factor in determining these findings, the results of this
meta-analysis need to be interpreted in light of the wide variability in the number of effect
sizes included for each risk factor, and the between-studies variability in effect sizes for some
of the risk factors.
Demographic and Pre-trauma Factors
Clear theoretically-driven predictions as to the impact of younger age on the risk of
developing PTSD have been difficult to make, given the often contradictory effects of age on
different processes underlying traumatic stress reactions (e.g. Meiser-Stedman, 2002;
Pfefferbaum, 1997; Salmon & Bryant, 2002). A clear picture emerged from our meta-analysis
concerning younger age as a risk factor for PTSD, with 18 studies included. The results of
this meta-analysis strongly suggest that overall younger age is largely unrelated to whether a
young person develops PTSD. Moderator analysis revealed that, unlike with female sex, there
was a statistically significantly stronger relationship between being younger and PTSD when
the index trauma was unintentional rather than intentional; however regardless of
intentionality, the population effect size estimate remained non significant. Similarly younger
age was found to have a different association with PTSD depending on whether the
potentially traumatic event was a group event or an individual one. Further analysis revealed
that it was a significant risk factor albeit with a small effect size (𝜌
̂ = 0.15), for group trauma
whereas it remained non significant for individual trauma. This is intriguing as it suggests

that if the event is experienced as part of a group, younger children may find it slightly harder
to process than older ones.
Only five of the 64 studies that were included in the analysis examined the
relationship between ethnicity and the development of PTSD. This is a smaller percentage of
studies than other reviews or meta-analyses such as the review by Foy et al. (1996) or the
meta-analysis of Brewin et al. (2000). In the present meta-analysis, ethnicity was found to
have a consistent but small effect on the likelihood of developing PTSD. As with the Brewin
et al. (2000) adult meta-analysis, it should be stressed that this variable was coded in a
dichotomous fashion (white versus black or minority ethnic), which may have masked
important differences between minority ethnic groups.
Low intelligence and low socio-economic status were more important, but still only
yielded small to medium effect sizes consistently reported across studies. Intelligence was
only examined in two studies and clearly needs further investigation.
A notable finding was the identification of female sex as a consistent, but ultimately
small (𝜌
̂= .15) risk factor for PTSD in children and adolescents. The direction of this effect
tallies with the conclusions of previous reviews (Davis & Siegel, 2000; Foy et al., 1996;
Pfefferbaum, 1997); however, this study helps to establish the limited extent of the
relationship. Placing strong emphasis on female sex in PTSD screening programs would
appear to be unwarranted. The observation of female sex as a significant if small risk factor
also concurs with data for adult populations (Brewin et al., 2000).
Moderator analysis found that female sex became a stronger risk factor in older
children and adolescents, and when the index trauma was unintentional. The age-related
effect may reflect the emergence of more differential responses to stressful events, e.g. the
tendency of female adolescents to use rumination (Hampel & Petermann, 2005; Broderick,
1998; Cutler & Nolen-Hoeksema, 1991; Nolen-Hoeksema & Girgus, 1994). That intentional

traumatic stressors should be related to a larger risk of PTSD in female youth, than other
types of trauma is an important observation. An increased sensitivity to interpersonal
violence in female youth may be understood from a variety of developmental theories (e.g.
neuroendocrinology, socialisation to particular sex roles), but presents an easily identifiable
specific type of trauma where this risk factor may be more important.
Pre-trauma factors (i.e. pre-trauma psychological problems, life events and parental
psychological problems) were also only small to medium risk factors. Objective trauma
characteristics (i.e. bereavement, time post-trauma, and trauma severity) also yielded only
small to medium effect sizes. Trauma severity (as indexed by objective criteria) was the
trauma characteristic most strongly associated with risk of developing PTSD. It was also
apparent, however, that this effect varied very widely. Clearly this is related strongly to the
very wide range of measures of trauma severity used, all of which are assumed to be directly
related to trauma severity.
Analysis revealed that none of the moderators that we investigated significantly
moderated the effect sizes of trauma severity. This might reflect the complexity of ‘trauma
severity’, the objective measurement of which is complicated by a number of conceptual
factors. For example, what is the essential property that makes a trauma traumatic; can
trauma exposure be adequately differentiated from trauma severity; do measurable properties
of a trauma correspond with the traumatic quality, so comparisons across trauma types are
likely to be even more complex. For example, one interpretation of the tendency of group
trauma to increase the link between trauma severity and risk of PTSD would be that in studies
where large groups of young people are exposed to trauma, trauma severity may be indexing
exposure (e.g. proximity to a natural disaster), rather than the severity of a trauma to which a
young person was clearly exposed (e.g. degree of injury following a motor vehicle accident).
Attempts to screen for children and adolescents at high risk of PTSD may therefore need to

index severity in a trauma-specific way, possibly limiting the applicability of some screening
tools.
Peritraumatic factors and meaning-making
Peri-traumatic fear and the perception of threat to life during the trauma had large
effect sizes as predictors of PTSD. Surprisingly, these variables had only been investigated
in 3 and 6 studies respectively, underlining the need for further research examining peri-
traumatic emotions and cognitions. While these data are clearly preliminary, these variables
are easily assessed using questionnaire measures and may prove to be of great utility in PTSD
screening tools for trauma-exposed youth (e.g. Winston, Kassam-Adams, Garcia-España,
Ittenbach & Cnaan, 2003; Nixon, Ellis, Nehmy & Ball 2010).
Individual Factors Post-trauma
A number of features of participants’ post-trauma coping and environment were
significant risk factors for PTSD. Blame others was found to have a medium to large effect
size, and all of the other individual post-trauma factors assessed (comorbid psychological
problems, distraction, PTSD at time 1, and thought suppression) yielded large effect sizes.
The presence of comorbid psychological problems, as with trauma severity, was a
significant risk factor for PTSD but with the degree of association varying considerably
across different studies. Moderation analysis revealed that anxiety, depression and other
psychological problems were all significant risk factors, depression was the most predictive
of PTSD symptoms. Furthermore, the presence of a comorbid psychological problem was
more of a risk factor for PTSD, in intentional trauma compared to unintentional trauma. The
wide variation for comorbid psychological problems likely reflects again the broad range of
traumas included within this meta-analysis, and the differing impacts that trauma can have on
young people. The way in which the interpersonal dimension of intentional trauma may lead
to wider psychiatric disturbance than unintentional trauma is an important issue. One possible

mechanism for this is greater cognitive misappraisal (for preliminary evidence of this see
Meiser-Stedman, Dalgleish, Glucksman, Yule & Smith, 2009). An alternative explanation is
that young people exposed to intentional trauma may be more likely to be living in toxic
environments, thus implicating pre-traumatic factors. Regardless of how comorbidity
precisely relates to PTSD, its importance as a risk factor for PTSD adds weight to two points
concerning the way in which clinicians consider young people’s responses to trauma. Firstly,
psychiatric reactions other than PTSD should be considered in the aftermath of trauma,
particularly certain types of trauma (e.g. intentional trauma). Secondly, understanding how
other psychiatric reactions (particularly depression) are elicited by trauma exposure is vital.
Work has begun to delineate the aetiology of different psychiatric responses in adults
exposed to trauma (e.g., Ehring, Ehlers & Glucksman, 2008), but no such work has been
conducted in youth.
Blame of others for the event was found to have a small to medium effect size, but
with only two studies being considered, further research is required before drawing firm
conclusions. However it is consistent with clinical experience where some clients are so
wrapped up in blaming someone for why an event has happened, they appear to be unable to
devote sufficient resources to processing what has happened. The emergence of distraction
and thought suppression as major risk factors for PTSD, albeit in only two studies each at this
stage, is significant, and these factors represent important targets for future investigation.
This relationship may merely reflect the overlap with the core avoidance symptoms of PTSD,
but, given the ease with which these psychological processes can be assessed, they may
nevertheless have a very important function within screening batteries. Indeed, these
variables compared favourably with the ability of an early index of PTSD to act as a risk
factor for PTSD one month post-trauma. Additionally, the fact that distraction and thought

suppression are amongst the largest effect sizes, supports the cognitive model’s assertion that
avoidance is central to the symptomatology.
The very large effect associated with T1 PTSD symptomatology as a risk factor for
later PTSD is hardly surprising, but this finding does help to emphasise the utility of
including early symptom levels within screening tools (e.g. the Trauma Screening
Questionnaire for adults; Brewin et al., 2002). Theoretically, these suggest that early
processes (i.e. either peri-trauma or in the immediate aftermath) play a major role in the
development of more significant and persistent PTSD symptoms.
Psychological Environmental Factors Post-trauma
All the features of the post-trauma psychological environment considered within this
meta-analysis showed small to medium or large effect sizes, indicating the strength of the
relationship between what happens after the event, with the development of PTSD. Life
events and parental psychological difficulties showed small to medium effect sizes whilst low
social support, family poor functioning and social withdrawal showed medium to large effect
sizes. While social support has been identified as being a factor of potential great importance
to the successful resolution of traumatic experience (e.g. Davis & Siegel, 2000; Pine &
Cohen, 2002), this meta-analysis only found 4 studies that have adequately addressed this
construct. Poor family functioning was a stronger risk factor for PTSD than poor parental
mental health, though parental mental health is the more established correlate of PTSD (being
examined in 25 studies), and its effect size was on the borderline between a medium and a
large effect size (p>0.3).
Limitations.
The present study is limited by the highly heterogeneous nature of studies in this area,
with the inclusion of papers that are widely ranging in terms of trauma type, assessment tools,
and the indices for several putative risk factors. A majority of the studies included in the

analysis used child-report measures of PTSD, with more studies it would be able to determine
whether this makes a difference to how the risk factors operate. This may be particularly
pertinent for younger children who may less able to report reliably on internal states. Future
meta-analyses and reviews may consider focusing on particular trauma types, in particular,
distinguishing between intended versus unintended and collective versus individual trauma
exposure (which accounted for much of the between-study variability in the present study).
Nevertheless, the PTSD diagnosis is not intended to be specific to any given form of
traumatic experience (within certain conceptual bounds), and between-trauma differences are
worth highlighting. The initial search was restricted to publications in English language, and
it is not known to what extent this limitation may have influenced the findings.
As already noted, many variables included in this meta-analysis have only been
examined in a handful of studies. Clearly this limits the conclusions that can be drawn about
these risk factors, but it does also help to highlight areas that warrant further investigation.
Implications
Despite the limitations outlined above, the present analysis does have some
implications for the theoretical understanding of PTSD in youth. First, the data concerning
pre-trauma factors shed light on the ultimate aetiology of PTSD and the diagnostic
distinctiveness of this construct. While the relationship between pre-trauma life events or
psychological problems and PTSD was significant across the studies that examined these
variables, the strength of this association was only modest in comparison to most of the peri-
trauma or post-trauma factors. This supports the suggestion that PTSD in youth is chiefly the
result of a reaction to the specific event, and as such is not particularly closely related to the
previous functioning of the young person, or to their previous experiences.
Second, the broadly consistent failure to observe a substantial relationship between
young age and PTSD needs to be acknowledged by clinicians and theoreticians alike.

Clinically, resources should not be devoted to younger children at the expense of adolescents.
Whether younger children are as able to report their distress and access help outside of
research studies is however a significant issue to be addressed in future studies. Theoretically,
this finding may suggest that developmental considerations are not as important as is often
thought. Developmental vulnerabilities for the onset and maintenance of PTSD may have
already been passed by the age of 8 years. Alternatively, a more complex picture may emerge
of young children being more vulnerable to developing PTSD because of some
developmentally sensitive factors, but more resilient due to other age-related factors (see
Salmon & Bryant, 2002, for an extensive review of how development may differentially
impact on the psychological processing of a potentially traumatic event). As the present meta-
analytic study was restricted to 6-18 year olds, it is not possible to say whether an age-related
vulnerability to developing PTSD is present or not in even younger children, though research
within this younger population is clearly more problematic (Scheeringa, Zeanah, Drell &
Larrieu, 1995).
Third, the strong relationship between PTSD risk and aspects of young people’s
recovery environment underlines the importance – and feasibility – of developing efficacious
post-trauma screening tools and inventions for children and adolescents. Demographic and
pre-trauma factors, are only weakly related to PTSD risk and are therefore unlikely to have
particular utility on their own as targets for screening tools (e.g. female gender, age, pre-
trauma psychological problems). Post-trauma factors, however, appear to be more strongly
implicated in the development of PTSD. Interventions based on these risk factors may
include empirically based screening tools for diverting resources to the most vulnerable
young people, and low intensity and cost effective interventions that do not rely on individual
psychological therapy. Preliminary work has begun to this end (e.g. Kenardy, Thompson, Le

Brocque & Olsson, 2008), but further research is required to specify which features of the
post-trauma recovery environment are actively toxic and which are epiphenomenal.
Some risk factors – namely, the perception of life threat, the use of distraction and
thought suppression – offer tentative support for a cognitive approach to understanding PTSD
in young people (e.g., Brewin, Dalgleish & Joseph, 1996; Dalgleish 1999, Ehlers & Clark,
2000; Meiser-Stedman, 2002). Such an account offers a framework for investigating the
specific mechanisms by which robust risk factors (e.g. parental psychological problems,
comorbidity, poor family functioning) may trigger or maintain post-traumatic stress in youth
(see Meiser-Stedman, 2002 for a fuller discussion), which would in turn inform the
development of more targeted and efficacious interventions. Regardless of how such a
cognitive research agenda turns out, a broadly psychosocial account of child and adolescent
PTSD would appear to be a paradigm for directing future research in this area.

References
References marked with an asterisk indicate studies included in the meta-analysis.
*Aaron, J., Zaglul, H., & Emery, R. (1999). Posttraumatic stress in children following acute
physical injury. Journal of Pediatric Psychology, 24, 335-343.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders (4th ed., text rev.). Washington, DC: Author.
*Asarnow, J., Glynn, S., Pynoos, R. S., Nahum, J., Guthrie, D., Cantwell, D. P., & Franklin,
B. (1999). When the earth stops shaking: Earthquake sequelae among children
diagnosed for pre-earthquake psychopathology. Journal of the American Academy of
Child and Adolescent Psychiatry, 38, 1016-1023.
*Bal, S., Oost, P. V., Bourdeaudhuij, I. D., & Crombez, G. (2003). Avoidant coping as a
mediator between self-reported sexual abuse and stress-related symptoms in
adolescents. Child Abuse & Neglect, 27, 883-897.
Begg, C. B. (1994). Publication bias. In H. Cooper & L. V. Hedges (Eds.), The Handbook of
Research Synthesis (pp. 399−409). New York: Russell Sage Foundation.
Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Gusman, F. D., Charney, D. S.,
& Keane, T. M. (1995). The development of a clinician-administered PTSD scale.
Journal of Traumatic Stress, 8, 75-90.
*Bradburn, I. S. (1991). After the earth shook: children’s stress symptoms 6-8 months after a
disaster. Advances in Behaviour Research & Therapy, 13, 173-179.
Brewin, C. R., Dalgleish, T., Joseph, S. (1996). A Dual Representation Theory of
Posttraumatic Stress Disorder Psychological Review, 103, 670-686.
Brewin, C. R., Andrews, B. & Valentine, J. D. (2000). Meta-analysis of risk factors for
posttraumatic stress disorder in trauma-exposed adults. Journal of Consulting and
Clinical Psychology, 68, 748-766

Brewin, C. R., Rose, S., Andrews, B., Green, J., Tata, P., McEvedy, C., Turner, S. & Foa, E.
B. (2002). Brief screening instrument for post-traumatic stress disorder. British
Journal of Psychiatry, 181, 158-162.
Broderick, P. C. (1998). Early adolescent gender differences in the use of ruminative and
distracting coping strategies. The Journal of Early Adolescence, 18, 173-191.
*Bryant, R. A., Salmon, K., Sinclair, E., & Davidson, P. (2007). Heart rate as a predictor of
posttraumatic stress disorder in children. General Hospital Psychiatry, 29, 66-68.
*Bryant, R. A., Salmon, K., Sinclair, E., & Davidson, P. (2007b). A prospective study of
appraisals in childhood posttraumatic stress disorder. Behaviour Research and
Therapy, 45, 2502–2507.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd
ed.). Hillsdale,
NJ: Erlbaum.
Cox, C. M., Kenardy,J. A. & Hendrikz, J. K. (2008). A meta-analysis of risk factors that
predict psychopathology following accidental trauma. Journal for Specialists in
Pediatric Nursing, 13, 98-110.
Cutler, S. E. & Nolen-Hoeksema, S. (1991). Accounting for sex differences in depression
through female victimization: Childhood sexual abuse. Sex Roles, 24, 425-438.
Dalgleish, T. (1999). Cognitive theories of post-traumatic stress disorder. In W. Yule (Ed.),
Post-traumatic stress disorders; Concepts and therapy. (pp. 193-220) New York,
John Wiley & Sons Ltd.
Dalgleish, T., Meiser-Stedman, R., & Smith, P. (2005). Cognitive aspects of posttraumatic
stress reactions and their treatment in children and adolescents: An empirical review
and some recommendations. Behavioural and Cognitive Psychotherapy, 33, 459–486.

Davis, L. & Siegel, L. J. (2000). Posttraumatic Stress Disorder in Children and Adolescents:
A Review and Analysis. Clinical Child and Family Psychology Review, 3(3), 135-
154.
De Bellis, M. D., Baum, A. S., Birmaher, B., Keshavan, M. S., Eccard, C. H., Boring, A. M.,
Jenkins, F. J, Ryan, N. D. (1999). Developmental traumatology: I. Biological stress
systems. Biological Psychiatry, 45, 1259-1270.
Dehon, C. & Scheeringa, M. (2006). Screening for preschool posttraumatic stress disorder
with the Child Behavior Checklist. Journal of Pediatric Psychology 31, 431–435.
*Delahanty, D. L., Nugent, N. R., Christopher, N. C., & Walsh, M. (2005). Initial urinary
epinephrine and cortisol levels predict acute PTSD symptoms in child trauma victims.
Psychoneuroendocrinology, 30, 121-128.
Dersimonian, R., & Laird, N. (1986). Meta-analysis in clinical-trials. Controlled Clinical
Trials, 7(3), 177−188.
Durlak, J. A., & Lipsey, M. W. (1991). A practitioner's guide to meta-analysis. American
Journal of Community Psychology, 19, 291−332.
Ehlers, A. & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder.
Behaviour Research & Therapy, 38, 319-345.
Ehring, T., Ehlers, A., & Glucksman, E. (2008). Do cognitive models help in predicting the
severity of posttraumatic stress disorder, phobia, and depression after motor vehicle
accidents? A prospective longitudinal study. Journal of Consulting and Clinical
Psychology, 76(2), 219-230.
Field, A. P. (2001). Meta-analysis of correlation coefficients: A Monte Carlo comparison of
fixed and random-effects methods. Psychological Methods, 6, 161–18.
Field, A. P. (2003). The problem in using fixed-effects models of meta-analysis on real-world
data. Understanding Statistics, 2, 77−96.

Field, A. P. (2005). Is the meta-analysis of correlation coefficients accurate when population
effect sizes vary? Psychological Methods, 10(4), 444−467.
Field, A. P. & Gillett, R. (2010). How to do a meta-analysis. British Journal of Mathematical
and Statistical Psychology, 63, 665–694.
Fletcher, K. E. (1996). Childhood posttraumatic stress disorder. In E. J. Mash and R. Barkley
(Eds.), Child Psychopathology (pp. 242–276). New York, NY: Guilford Press.
Foy, D. W., Madvig, B. T., Pynoos, R. S. & Camilleri, A. J. (1996). Etiologic Factors in the
Development of Posttraumatic Stress Disorder in Children and Adolescents. Journal
of School Psychology 34(2), 133-145.
*Giannopoulou, I., Strouthos, M., Smith, P., Dikaiakou, A., Galanopoulou, V., & Yule, W.
(2006). Post-traumatic stress reactions of children and adolescents exposed to the
Athens 1999 earthquake. European Psychiatry, 21, 160-166.
*Goenjian, A. K., Molina, L., Steinberg, A. M., Fairbanks, L. A., Alvarez, M. L., Goenjian,
H. A., & Pynoos, R. S. (2001). Posttraumatic stress and depressive reactions among
Nicaraguan adolescents after Hurricane Mitch. American Journal of Psychiatry, 158,
788-794.
*Goenjian, A. K., Pynoos, R., Steinberg, A. M., Najarian, L. M., Asarnow, J., Karayan, I.,
Ghurabi, M., & Fairbanks, L. A. (1995). Psychiatric comorbidity in children after the
1988 earthquake in Armenia. Journal of the American Academy of Child and
Adolescent Psychiatry, 34, 1174-1184.
*Goenjian, A. K., Walling, D., Steinberg, A. M., Karayan, I., Najarian, L. M., & Pynoos, R.
(2005). A prospective study of posttraumatic stress and depressive reactions among
treated and untreated adolescents 5 years after a catastrophic disaster. American

*Goenjian, A. K., Walling, D., Steinberg, A. M., Roussos, A., Goenjian, H. A., & Pynoos, R.
(2009). Depression and PTSD symptoms among bereaved adolescents 6½ years after
the 1988 spitak earthquake. Journal of Affective Disorders, 112, 81–84.
Hampel, P. & Petermann, F. (2005). Age and Gender Effects on Coping in Children and
Adolescents. Journal of Youth and Adolescence, 34, 73-83.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL:
Academic Press.
Hedges, L. V., & Vevea, J. L. (1998). Fixed- and random-effects models in meta-analysis.
Psychological Methods, 3, 486−504.
*Heptinstall, E., Sethna, V., & Taylor, E. (2004). PTSD and depression in refugee children:
Associations with pre-migration trauma and post-migration stress. European Child
and Adolescent Psychiatry, 13, 373-380.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of Meta-analysis: correcting error and bias
in research findings. Newbury Park, CA: Sage.
Hunter, J. E., & Schmidt, F. L. (2000). Fixed effects vs. random effects meta-analysis
models: Implications for cumulative research knowledge. International Journal of
Selection and Assessment, 8, 275−292.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias
in research findings (2nd ed.). Newbury Park, CA: Sage.
*Husain, S. A., Nair, J., Holcomb, W., Reid, J. C., Vargas, V., & Nair, S. S. (1998). Stress
reactions of children and adolescents in war and siege conditions. American Journal
of Psychiatry, 155, 1718-1719.
*Jaycox, L. H., Styein, B. D., Kataoka, S. H., Wong, M., Fink, A., Escudero, P., & Zaragoza,
C. (2002). Violence exposure, posttraumatic stress disorder, and depressive symptoms

among recent immigrant schoolchildren. Journal of the American Academy of Child
and Adolescent Psychiatry, 41, 1104-1110.
Kenardy, J., Thompson, K., Le Brocque, R., & Olsson, K. (2008). Information-provision
intervention for children and their parents following pediatric accidental injury.
European Child & Adolescent Psychiatry, 17, 316-325.
*Kilic, Z. E., Ozguven, H. D., & Sayil, I. (2003). The psychological effects of parental
mental health on children experiencing disaster: The experience of Bolu earthquake in
Turkey. Family Process, 42, 485-495.
*Koplewicz, H. S., Vogel, J. M., Solanto, M. V., Morrissey, R. F., Alonso, C. M., Abikoff,
H., Gallagher, R., & Novick, R. M. (2002). Child and parent response to the 1993
World Trade Centre bombing. Journal of Traumatic Stress, 15, 77-85.
*Korol, M., Green, B. L., & Gleser, G. L. (1999). Children’s responses to a nuclear waste
disaster: PTSD symptoms and outcome prediction. Journal of the American Academy
of Child and Adolescent Psychiatry, 38, 368-375.
*La Greca, A. M., Silverman, W. K., Vernberg, E. M., & Prinstein, M. J. (1996). Symptoms
of posttraumatic stress in children after Hurricane Andrew: A Prospective study.
Journal of Consulting and Clinical Psychology, 64, 712-723.
*Lack, C. W., & Sullivan, M. A. (2008). Attributions, coping, and exposure as predictors of
long-term posttraumatic distress in tornado-exposed children. Journal of Loss and
Trauma, 13, 72–84.
*Landolt, M. A., Vollrath, M., Ribi, K., Gnehm, H. E., & Sennhauser, F. H. (2003).
Incidence and associations of parental and child posttraumatic stress symptoms in
pediatric patients. Journal of Child Psychology and Psychiatry, 44(8), 1199-1207.
*Landolt, M. A., Vollrath, M., Timm, K., Gnehm, H. E., & Sennhauser, F. H. (2005).
Predicting posttraumatic stress symptoms in children after road traffic accidents.

Journal of the American Academy of Child and Adolescent Psychiatry, 44, 1276-
1283.
*Laor, N., Wolmer, L., Kora, M., Yucel D., Spirman, S., & Yazgan, Y. (2002).
Posttraumatic, dissociative and grief symptoms in Turkish children exposed to the
1999 earthquake. The Journal of Nervous and Mental Disease, 190, 824-832.
*Lengua, L. J., Long, A. C., Smith, k. I., & Meltzoff, A. N. (2005). Pre-attack
symptomatology and temperament as predictors of children’s responses to the
September 11 terrorist attacks. Journal of Child Psychology and Psychiatry, 46, 631-
645.
*Lengua, L. J., Long, A. C., & Meltzoff, A. N. (2006). Pre-attack stress-load, appraisals, and
coping in children’s responses to the 9/11 terrorist attacks. Journal of Child
Psychology and Psychiatry, 47, 1219-1227.
Light, R. J., & Pillemer, D. B. (1984). Summing up: The science of reviewing research.
Cambridge, MA: Harvard University Press.
*Linning, L. M., & Kearney, C. A. (2004). Post-traumatic stress disorder in maltreated youth:
A study of diagnostic comorbidity and child factors. Journal of Interpersonal
Violence, 19, 1087-1101.vn
*Lonigan, C. J., Shannon, M. P., Finch, A. J. (Jr)., Daugherty, T. K., & Taylor, C. M. (1991).
Children’s reactions to a natural disaster: Symptom severity and degree of exposure.
Advances in Behaviour Research and Therapy, 13, 135-154.
*Magal-Vardi, O., Laor, N., Toren, A., Strauss, L., Wolmer, L., Bielorai, B., Rechavi, G., &
Toren P. (2004). Psychiatric morbidity and quality of life in children with
malignancies and their parents. Journal of Nervous & Mental Disease, 192, 872-875.

*Max, J. E., Castillo, C. S., Robin, D. A., Lindgren, S. D., Smith, W. L., Sato, Y., & Arndt,
S. (1998). Posttraumatic stress symptomatology after childhood traumatic brain
injury. The Journal of Nervous & Mental Disease, 186, 589-596.
*McDermott, B. M., Lee, E. M., Judd, M., & Gibbon, P. (2005). Posttraumatic stress disorder
and general psychopathology in children and adolescents following a wildfire
disaster. Canadian Journal of Psychiatry, 50, 137-143.
Meiser-Stedman, R. (2002). Towards a Cognitive–Behavioral Model of PTSD in Children
and Adolescents. Clinical Child and Family Psychology Review, 5, 217-232.
Meiser-Stedman, R., Dalgleish, T., Glucksman, E., Yule, W., Smith, P. (2009). Maladaptive
cognitive appraisals mediate the evolution of posttraumatic stress reactions: A 6-
month follow-up of child and adolescent assault and motor vehicle accident survivors.
Journal of Abnormal Psychology, 118, 778-787.
*Meiser-Stedman, R. A., Yule, W., Dalgleish, T., Smith, P., & Glucksman, E. (2005). The
role of the family in child and adolescent posttraumatic stress following attendance at
an emergency department. Journal of Pediatric Psychology, 31(1), 1-6.
Meiser-Stedman, R., Smith, P., Glucksman, E., Yule, W., & Dalgleish, T. (2008). The
posttraumatic stress disorder diagnosis in preschool- and elementary school-age
children exposed to motor vehicle accidents. American Journal of Psychiatry,
165(10), 1326-1337.
*Mirza, K. A. H., Bhadrinath, I. M., Goodyear, I. M., & Gilmour, C. (1998). Post-traumatic
stress disorder in children and adolescents following road traffic accidents. British
National Institute for Health & Clinical Excellence. (2005). Post-traumatic stress disorder:
the management of PTSD in adults and children in primary and secondary care.
London: Gaskell and the British Psychological Society.

National Research Council (1992). Combining information: Statistical issues and
opportunities for research. Washington, D.C.: National Academy Press.
Nixon, R. D. V., Ellis, A. A., Nehmy, T. J., Ball, S-A. (2010). Screening and predicting
posttraumatic stress and depression in children following single-incident trauma.
Journal of Clinical Child and Adolescent Psychology, 39, 588-596.
Nolen-Hoeksema, S. & Girgus, J. S. (1994). The emergence of gender differences in
depression during adolescence. Psychological Bulletin, 115, 424-443
*Nugent, N. R., Christopher, N. C., & Delahanty, D. L. (2006). Initial physiological
responses and perceived hyperarousal predict subsequent emotional numbing in
pediatric injury patients. Journal of Traumatic Stress, 19, 349-359.
*Nugent, N. R., Ostrowski, S., Christopher, N. C., & Delahanty, D. L. (2007). Parental
posttraumatic stress symptoms as a moderator of child’s acute biological response and
subsequent posttraumatic stress symptoms in pediatric injury patients. Journal of
Pediatric Psychology, 32, 309-318.
Osburn, H. G., & Callender, J. (1992). A note on the sampling variance of the mean
uncorrected correlation in meta-analysis and validity generalization. Journal of
Applied Psychology, 77, 115-122.
*Ostrowski, S. A., Christopher, N. C., & Delahanty, D. L. (2007). The impact of maternal
posttraumatic stress disorder symptoms and child gender on risk for persistent
posttraumatic stress disorder symptoms in child trauma victims. Journal of Pediatric
Psychology, 32, 338-342.
*Ostrowski, S. A., Christopher, N. C., & Delahanty, D. L. (2007b). Acute child and mother
psychophysiological responses and subsequent PTSD symptoms following a child’s
traumatic event. Journal of Traumatic Stress, 20, 677–687.

*Otto, M. W., Henin, A., Hirshfeld-Becker, D. R., Pollack, M. H., Biederman, J., &
Rosenbaum, J. F. (2007). Posttraumatic stress disorder symptoms following media
exposure to tragic events: Impact of 9/11 on children at risk for anxiety disorders.
Journal of Anxiety Disorders, 21, 888–902.
*Overstreet, S., Dempsey, M., Graham, D., & Moely, B. (1999). Availability of family
support as a moderator of exposure to community violence. Journal of Clinical Child
Overton, R. C. (1998). A comparison of fixed-effects and mixed (random-effects) models for
meta-analysis tests of moderator variable effects. Psychological Methods, 3(3),
354−379.
Ozer, E. J., Best, S. R., Lipsey, T. L. & Weis, D. S. (2003). Predictors of Posttraumatic Stress
Disorder and Symptoms in Adults: A Meta-Analysis. Psychological Bulletin, 129,
52–73.
*Pelcovitz, D., Kaplan, S. J., DeRosa, R. R., Mandel, F. S., & Salzinger, S. (2000).
Psychiatric disorders in adolescents exposed to domestic violence and physical abuse.
American Journal of Orthopsychiatry, 70, 360-369.
Perrin, S., Smith, P. & Yule, W. (2000). Practitioner review: The assessment and treatment of
Post-traumatic Stress Disorder in children and adolescents. Journal of Child
Psychology and Psychiatry ,41,277-289.
Pfefferbaum, B. (1997). Posttraumatic stress disorder in children: A review of the past 10
years. Journal of the American Academy of Child and Adolescent Psychiatry, 36(11),
1503-1511.
*Pfefferbaum, B., Nixon, S. J., Tucker, P. M., Tivis, R. D., Moore, V. L., Gurwitch, R. H.,
Pynoos, R. S., & Geis, H, K. (1999). Posstraumatic stress responses in bereaved

children after the Oklahoma City bombing. Journal of the American Academy of
Child and Adolescent Psychiatry, 38, 1372-1379.
*Pfefferbaum, B., Seale, T. W., McDonald, N. B., Brandt, E. N., Rainwater, S. M. Jr.,
Maynard, B. T., Meiehoefer, B., & Miller, P. D. (2000). Posttraumatic stress two
years after the Oklahoma city bombing in youths geographically distant from the
explosion. Psychiatry, 63(4), 358-370.
*Pfefferbaum, B. (2001). The impact of the Oklahoma City bombing on children in the
community. Military Medicine, 166, 49-50.
Pine, D. S. & Cohen, J. A. (2002). Trauma in Children and Adolescents: Risk and Treatment
of Psychiatric Sequelae. Biological Psychiatry, 51, 519-531.
*Punamaki, R-L., Qouta, S., Sarraj, E. E., & Montgomery, E. (2006). Psychological distress
and resources among siblings and parents exposed to traumatic events. International
Journal of Behavioral Development, 30, 385-397.
Pynoos, R. S., Steinberg, A. M. & Piacentini, J. C. (1999). A developmental psychopathology
model of childhood traumatic stress and intersection with anxiety disorders.
Biological Psychiatry, 46, 1542-1554.
R Development Core Team (2010). R: A language and environment for statistical computing.
Vienna, Austria: R Foundation for Statistical Computing.
Rosenthal, R. (1979). The ‘File Drawer Problem’ and tolerance for null results. Psychological
Bulletin, 86, 638−641.
Rosenthal, R. (1991). Meta-analytic procedures for social research (revised ed.). Newbury
Park, CA: Sage.
Rosenthal, R. (1995). Writing meta-analytic reviews. Psychological Bulletin, 118, 183−192.
Rosenthal, R., & DiMatteo, M. R. (2001). Meta-analysis: Recent developments in
quantitative methods for literature reviews. Annual Review of Psychology, 52, 59−82.

*Roussos, A., Goenjian, A. K., Steinberg, A. M., Sotiropoulou, C., Kakaki, M., Kabakos, C.,
Karagianni, S., & Manouras, V. (2005). Posttraumatic stress and depressive reactions
among children and adolescents after the 1999 earthquake in Ano Liosia, Greece.
American Journal of Psychiatry, 162, 530-537.
*Runyon, M. K., & Kenny, M. C. (2002). Relationship of attributional style, depression, and
posttrauma distress among children who suffered physical or sexual abuse. Child
Maltreatment, 7, 254-264.
Salmon, K. & Bryant, R. (2002). Posttraumatic stress disorder in children. The influence of
developmental factors. Clinical Psychology Review, 22, 163-188.
*Saltzman, K. M., Weems, C. F., & Carrion, V. G. (2006). IQ and posttraumatic stress
symptoms in children exposed to interpersonal violence. Child Psychiatry and
Human Development, 36, 261-272.
*Saxe, G. N., Stoddard, F., Hall, E., Chawla, N., Lopez, C., Sheridan, R., King, D., King, L.,
Yehuda, R. (2005). Pathways to PTSD, part I: Children with burns. American Journal
of Psychiatry, 162, 1299-1304.
*Schafer, I., Barkmann, C., Riedesser, P., Schulte-Markwort, M. (2004). Peritraumatic
dissociation predicts posttraumatic stress in children and adolescents following road-
traffic accidents. Journal of Trauma & Dissociation, 5, 79-92.
*Schafer, I., Barkmann, C., Riedesser, P., Schulte-Markwort, M. (2006). Posttraumatic
syndromes in children and adolescents after road traffic accidents – A prospective
cohort study. Psychopathology, 39, 159-164.
Scheeringa M. S. (2003). New findings on alternative criteria for PTSD in preschool children.
Journal of the American Academy of Child and Adolescent Psychiatry, 42, 561-570

Scheeringa, M. S., Zeanah,C. H., Drell, M. J. & Larrieu, J. A. (1995). Two approaches to the
diagnosis of posttraumatic stress disorder in infancy and early childhood. Journal of
the American Academy of Child and Adolescent Psychiatry, 34, 191-200.
*Shannon, M. P., Lonigan, C. J., Finch, A. J. (Jr)., & Taylor, C. M. (1994). Children exposed
to disaster: I. Epidemiology of post-traumatic symptoms and symptom profiles.
Journal of the American Academy of Child and Adolescent Psychiatry, 33, 80-93.
*Shemesh, E., Newcorn, J. H., Rockmore, L., Shneider, B. L., Emre, S., Gelb, B. D.,
Rappaport, R., Noone, S. A., Annunziato, R., Schmeidler. S., & Yehuda, R. (2005).
Comparison of parent and child reports of emotional trauma symptoms in pediatric
outpatient settings. Pediatrics, 115, 582-589.
*Smith, S., Perrin. S., Yule, W., & Rabe-Hesketh, S. (2001). War exposure and maternal
reactions in the psychological adjustment of children from Bosnia-Hercegovina.
Journal of Child Psychology & Psychiatry, 42(3), 395-404.
*Solomon, Z., & Lavi, T. (2005). Israeli youths in the second intifada: PTSD and future
orientation. Journal of the American Academy of Child and Adolescent Psychiatry,
44, 1167-1175.
*Stallard, P., & Smith, E. (2007). Appraisals and cognitive coping styles associated with
chronic post-traumatic symptoms in child road traffic accident survivors. Journal of
Child Psychology and Psychiatry, 48, 194-201.
*Stallard, P., Salter, E., & Velleman, R. (2004). Post-traumatic stress disorder following road
traffic accidents: A second prospective study. European Child and Adolescent
Psychiatry, 13, 172-178.
*Stallard, P., Velleman, R., Langsford, J., & Baldwin, S. (2001). Coping and psychological
distress in children involved in road traffic accidents. British Journal of Clinical

Steinberg, A. M., Brymer, M. J., Decker, K. B. & Pynoos, R.S. (2004). The University of
California at Los Angeles Post-traumatic Stress Disorder Reaction Index. Current
Psychiatry Reports, 6, 96–100.
*Thabet, A. A., Abed, Y., & Vostanis, P. (2002). Emotional problems in Palestinian children
living in a war zone: A cross-sectional study. The Lancet, 25, 1801-1804.
*Thabet, A. A., & Vostanis, P. (1999). Posttraumatic stress reactions in children of war.
Journal of Child Psychology and Psychiatry, 40, 385–391.
*Thabet, A. A. M., Abed, Y., & Vostanis, P. (2004). Comorbidity of PTSD and depression
among refugee children during war conflict. Journal of Child Psychology and
Psychiatry, 45, 533-542.
*Udwin, O., Boyle, S., Yule, W., Bolton, D., & O’Ryan, D. (2000). Risk factors for long-
term psychological effects of a disaster experienced in adolescence: Predictors of post
traumatic stress disorder. Journal of Child Psychology & Psychiatry, 41, 969-979.
*Vernberg, E. M., LaGreca, A. M., Silverman, W. K., & Prinstein, M. J. (1996). Prediction of
posttraumatic stress symptoms in children after Hurricane Andrew. Journal of
Abnormal Psychology, 105(2), 237-248.
Vevea, J. L., & Woods, C. M. (2005). Publication bias in research synthesis: Sensitivity
analysis using a priori weight functions. Psychological Methods, 10, 428−443.
*Wickrama, K. A. S., & Kaspar, V. (2007). Family context of mental health risk in Tsunami-
exposed adolescents: Findings from a pilot study in Sri Lanka. Social Science &
Medicine, 64, 713-723.
Winston, F. K., Kassam-Adams, N., Garcia-España, F., Ittenbach, R., Cnaan, A. (2003).
Screening for Risk of Persistent Posttraumatic Stress in Injured Children and Their
Parents. Journal of the American Medical Association, 290, 643-649.

Table 1
Risk Factors Extracted From Included Studies and Descriptive Statistics of the Overall Effect Sizes From Each Study (different psychological
problems were combined for analysis)
Risk Factor Article Name Assessment of risk factor k Mean r SD Min. Max.
Bereavement Goenjian et al., 1995 Death of family members 3 .19 .04 .14 .22
Lengua et al., 2005 Knowing someone that died 1 .16
Pfefferbaum et al., 1999 Knowing someone killed 3 .16 .03 .12 .19
Wickrama & Kaspar, 2007 Death of someone close 2 .35 .01 .34 .35
Blame others Stallard et al., 2001 Kidcope 1 .29
Vernberg et al., 1996 Kidcope 1 .56
Comorbid Anxiety Goenjian et al., 1995 Diagnostic Interview Schedule for Children
and Adolescents
1 .19
Kilic et al., 2003 State-Trait Anxiety Inventory for Children 2 .22 .03 .20 .24
Lack & Sullivan, 2008 Tornado Exposure Questionnaire 1 .24
Lengua et al., 2005 Revised Children’s Manifest Anxiety Scale 1 .69
Linning & Kearney, 2004 Anxiety Disorders Interview Schedule for
Children – Child Version
3 .24 .01 .24 .25
Lonigan et al., 1991 Revised Children’s Manifest Anxiety Scale 2 .41 .20 .27 .55
Saxe et al., 2005 Multidimensional Anxiety Scale for
Children
1 .68
Schafer et al., 2006 Anxiety Disorders Interview Schedule –
Revised
1 .21
Udwin et al., 2000 Revised Children’s Manifest Anxiety Scale 3 .35 .12 .22 .45
Comorbid conduct
problem
Lengua et al., 2005 Youth Self Report – Delinquent &
Aggressive Behaviour Subscales
1 .35
Comorbid Depression Goenjian et al., 1995 Depression Self Rating Scale 4 .42 .11 .28 .55
Goenjian et al., 2001 Depression Self Rating Scale 4 .68 .11 .56 .81
Jaycox et al., 2002 Children’s Depression Inventory 1 .54

Kilic et al., 2003 Children’s Depression Inventory 1 .10
Lengua et al., 2005 Children’s Depression Inventory 1 .46
Linning & Kearney, 2004 Anxiety Disorders Interview Schedule for
Children – Child Version
2 .41 .04 .38 .44
Overstreet et al., 1999 Children’s Depression Inventory 1 .18
Roussos et al., 2005 Depression Self Rating Scale 3 .54 .03 .51 .56
Schafer et al., 2006 Children’s Depression Inventory 1 .22
Thabet et al., 2004 Short Mood and Feelings Questionnaire 2 .34 .11 .26 .42
Udwin et al., 2000 Depression Self Rating Scale 3 .37 .10 .29 .48
Wickrama & Kaspar, 2007 Youth Self Report – depression items 1 .54
Comorbid internalising Otto et al, 2007 Behavioural Inhibition 2 .12 .05 .08 .15
Comorbid Psychological
Problem
McDermott et al., 2005 Strength and Difficulties Questionnaire 1 .22
Udwin et al., 2000 Child and Adolescent Psychopathology
Interview
2 .18 .03 .16 .20
Distraction Stallard & Smith, 2007 Two questions to assess cognitive coping
strategies
1 .58
Stallard et al., 2001 Kidcope 2 .29 .01 .28 .30
Female Gender Delahanty et al., 2005 Standard demographic information 1 .36
Goenjian et al., 1995 Standard demographic information 4 .14 .15 −.07 .26
Goenjian et al., 2001 Standard demographic information 4 .20 .09 .07 .28
Goenjian et al., 2005 Standard demographic information 2 .21 .12 .13 .29
Heptinstall et al., 2004 Standard demographic information 1 −.18
Husain et al., 1998 Standard demographic information 1 .15
Jaycox et al., 2002 Standard demographic information 1 .09
Kilic et al., 2003 Standard demographic information 1 .36
Korol et al., 1999 Standard demographic information 1 .06
Landolt et al., 2003 Standard demographic information 1 .12
Landolt et al., 2005 Standard demographic information 2 .05 .06 .01 .09
Lengua et al., 2005 Standard demographic information 1 .05
Linning & Kearney, 2004 Standard demographic information 1 .35

Lonigan et al., 1991 Standard demographic information 1 .15
McDermott et al., 2005 Standard demographic information 1 .07
Mirza et al., 1998 Standard demographic information 4 .27 .03 .22 .29
Ostrowski et al, 2007 Standard demographic information 2 .24 .07 .18 .29
Otto et al, 2007 Standard demographic information 2 .11 .03 .09 .14
Overstreet et al., 1999 Standard demographic information 1 .14
Pfefferbaum et al., 2000 Standard demographic information 1 .21
Roussos et al., 2005 Standard demographic information 1 .16
Runyon & Kenny, 2002 Standard demographic information 1 .19
Schafer et al., 2004 Standard demographic information 1 .25
Shannon et al., 1994 Standard demographic information 1 .20
Stallard & Smith, 2007 Standard demographic information 1 .29
Stallard et al., 2004 Standard demographic information 1 .30
Thabet et al., 2004 Standard demographic information 1 .01
Udwin et al., 2000 Standard demographic information 1 .18
Vernberg et al., 1996 Standard demographic information 1 .10
Low Intelligence Saltzman et al., 2006 Wechsler Abbreviated Scales of Intelligence 9 .25 .11 .12 .40
Udwin et al., 2000 Presence of learning difficulties 2 .18 .00 .18 .18
Low SES Delahanty et al., 2005 Parental income 1 .26
Landolt et al., 2003 Paternal occupation & maternal education 1 .06
Landolt et al., 2005 Paternal occupation & maternal education 2 .03 .01 .02 .03
Max et al., 1998 Hollingshead Four Factor Index 4 .27 .05 .20 .30
Nugent et al., 2007 Parental Education 2 .16 .09 .10 .22
Ostrowski et al, 2007b Parental Income 1 .48
Otto et al, 2007 Hollingshead Four Factor Index 2 .05 .04 .02 .08
Low Social Support Heptinstall et al., 2004 Feeling isolated or excluded 2 −.15 .15 −.26 −.04
Stallard & Smith, 2007 Perceived alienation 1 .59
Udwin et al., 2000 Social Support Scale
Social Relationships Problems
3 .29 .08 .22 .38

Vernberg et al., 1996 Social Support Scale for Children and
Adolescents
5 .35 .22 .20 .73
Maternal Anxiety Smith et al., 2001 State – Trait Anxiety Inventory 1 .24
Maternal Depression Kilic et al., 2003 Beck Depression Inventory 1 .19
Smith et al., 2001 Beck Depression Inventory 1 .17
Wickrama & Kaspar, 2007 Center for Epidemiological Studies
Depression Scale
1 .23
Maternal Psychological
Problem
Kilic et al., 2003 General Health Questionnaire 1 .26
Smith et al., 2001 General Health Questionnaire 1 .19
Maternal PTSD Landolt et al., 2003 Posttraumatic Diagnostic Scale 1 −.01
Landolt et al., 2005 Posttraumatic Diagnostic Scale 2 .15 .13 .06 .24
Ostrowski et al, 2007 Clinician Administered PTSD Scale 3 .37 .16 .23 .55
Punamaki et al., 2006 PTSD Reaction Index 3 .11 .11 .00 .22
Smith et al., 2001 Impact of Events Scale – Revised 2 .35 .03 .33 .37
Wickrama & Kaspar, 2007 DSM-IV diagnostic
interview items
1 .40
Parental Depression Meiser-Stedman et al., 2005 Beck Depression Inventory 2 .42 .03 .40 .44
Parental psychological
problem
Linning & Kearney, 2004 Family Environment Scale 1 .28
Nugent et al., 2007 Symptom Checklist – 90 revised 1 .50
Parental PTSD Koplewicz et al., 2002 Adult Posttraumatic Stress Reaction Index 4 .48 .30 .08 .81
Landolt et al., 2005 Posttraumatic Diagnostic Scale 1 .52
Magal-Vardi et al., 2004 Davidson Trauma Scale 2 .55 .11 .47 .63
Nugent et al., 2007 Impact of Events Scale – Revised 3 .38 .13 .23 .45
Shemesh et al., 2005 Impact of Events Scale 2 .33 .01 .32 .34
Paternal Depression Kilic et al., 2003 Beck Depression Inventory 1 .58
Paternal Psychological
Problem
Kilic et al., 2003 General Health Questionnaire 1 .52
Paternal PTSD Landolt et al., 2003 Posttraumatic Diagnostic Scale 1 .02
Landolt et al., 2005 Posttraumatic Diagnostic Scale 2 .16 .21 .01 .31
Punamaki et al., 2006 Adult
Posttraumatic Stress Reaction Index
3 .15 .10 .10 .26

Perceived Life Threat Lack & Sullivan, 2008 Trauma Exposure Questionnaire – Children 1 .41
Landolt et al., 2005 Threat appraisal question 2 .33 .06 .29 .37
McDermott et al., 2005 Threat appraisal question 1 .40
Solomon & Lavi, 2005 Threat appraisal question 3 .31 .12 .17 .39
Udwin et al., 2000 Threat appraisal questions 2 .27 .06 .22 .31
Vernberg et al., 1996 Hurricane-Related Traumatic Experiences
Questionnaire
1 .39
Peri-trauma fear Aaron et al., 1999 Likert scale question 1 .56
Lack & Sullivan, 2008 Trauma Exposure Questionnaire 2 .36 .17 .24 .48
Udwin et al., 2000 Fear and panic questions 3 .18 .06 .12 .24
Poor Family Functioning Kilic et al., 2003 McMaster Family Assessment Device 7 .07 .09 −.03 .21
Meiser-Stedman et al., 2005 Family Functioning Questionnaire 1 .43
Otto et al, 2007 Parental Expressed Emotion (criticism or
over-involvement)
4 .05 .02 .03 .07
Pelcovitz et al., 2000 Parental Bonding Instrument,
Family Adaptability and Cohesion
Evaluation Scale III,
Family Disagreements Interview
4 .78 .10 .65 .87
Roussos et al., 2005 Difficulties at home 1 .31
Udwin et al., 2000 Child & Adolescent Psychopathology
Interview (violence at home)
1 .15
Wickrama & Kaspar, 2007 Parent-child relationship quality 1 .79
Post-Trauma Life Events Heptinstall et al., 2004 Uncertain asylum application 7 .17 .18 −.03 .49
Landolt et al., 2005 Occurrence of 12 major life events 2 −.15 .08 −.20 −.09
McDermott et al., 2005 Having to live elsewhere 1 .06
Udwin et al., 2000 Life Events Scale 3 .34 .08 .28 .44
Wickrama & Kaspar, 2007 Number of days displaced from home 1 .29
Post-Trauma Media
Exposure
Lengua et al., 2006 Parental responses to questions about child
media exposure
1 −.16
Otto et al, 2007 Parental responses to questions about child
media exposure
4 .11 .05 .05 .17

Pfefferbaum et al., 2000 Child responses to questions about media
exposure
3 .37 .04 .34 .41
Pre-Trauma Anxiety Asarnow et al., 1999 Revised Children’s Manifest Anxiety Scale 2 .19 .01 .19 .20
Lengua et al., 2005 Child Behaviour Checklist,
Revised Children’s Manifest Anxiety Scale
3 .08 .03 .05 .11
Mirza et al., 1998 Revised Children’s Manifest Anxiety Scale 2 .20 .02 .18 .21
Otto et al, 2007 Reported history of child anxiety disorder 2 .07 .02 .06 .08
Udwin et al., 2000 Retrospective Experiences and Child and
Adolescent Psychopathology Interview
1 .17
Pre-trauma Conduct
Problem
Lengua et al., 2005 Child Behaviour Checklist and Achenbach
Youth Self Report – Delinquent and
Aggressive behavior subscales
3 .14 .09 .05 .22
Mirza et al., 1998 Young Person's Questionnaire (Parent
Version) - Behaviour Indicator Scale
Subscale
2 .16 .19 .02 .29
Pre-Trauma Depression Lengua et al., 2005 Child Behaviour Checklist,
Child Depression Inventory
3 .14 .12 .00 .22
Mirza et al., 1998 Mood and Feelings Questionnaire (Parent
report)
2 .21 .00 .21 .21
Otto et al, 2007 Kiddie Schedule for Affective Disorders and
Schizophrenia, Epidemiologic Version
(KSADS-E)
2 .07 .01 .07 .08
Pre-trauma family
psychological problem
Max et al., 1998 Family History Research Diagnostic Criteria 2 .14 .18 .01 .27
Pre-trauma internalising Aaron et al., 1999 Child Behaviour Checklist 1 .39
Pre-Trauma Life Events Heptinstall et al., 2004 Number of pre-migration traumas 2 .48 .09 .41 .54
Landolt et al., 2003 Recent significant life events 1 .05
Landolt et al., 2005 Occurrence of major life events 2 .20 .08 .14 .25
Lengua et al., 2006 General Life Events Schedule for Children 1 .38
Overstreet et al., 1999 Life Events Checklist 1 .16

Stallard & Smith, 2007 Previous RTA,
Previous accident
2 .16 .04 .13 .19
Stallard et al., 2004 Previous RTA 1 .10
Udwin et al., 2000 Medical Problems 1 .23
Pre-trauma low self-
esteem
Lengua et al., 2005 Perceived Competence Scale for Children-
Self-worth Subscale
2 .18 .04 .15 .21
Mirza et al., 1998 Parent's version of Young Persons'
Questionnaire.
2 .14 .10 .08 .21
Pre-trauma parental
anxiety
Otto et al, 2007 Structured Clinical Interview
for DSM-III-R (SCID)
2 .06 .03 .04 .09
Pre-trauma parental
depression
Otto et al, 2007 Structured Clinical Interview
for DSM-III-R (SCID)
2 .15 .09 .08 .21
Pre-trauma parental
psychological problem
Max et al., 1998 Family History Research Diagnostic Criteria 2 .22 .03 .20 .24
Pre-Trauma
Psychological Problem
Max et al., 1998 Pre-injury internalizing disorder 4 .34 .04 .30 .38
Mirza et al., 1998 Rutter’s Questionnaire (parent) 1 .11
Udwin et al., 2000 Retrospective Experiences and Child and
Adolescent Psychopathology Interview
1 .16
PTSD (T1) Koplewicz et al., 2002 Child PTSD Reaction Index 2 .79 .01 .78 .79
Landolt et al., 2005 Child PTSD Reaction Index 2 .31 .47 −.02 .64
Nugent et al., 2007 Clinician Administered PTSD Scale – Child
and Adolescent
1 .75
Udwin et al., 2000 Impact of Events Scale 1 .22
Race (BME) La Greca et al., 1996 Standard demographic information 6 .20 .04 .15 .25
Lengua et al., 2005 Standard demographic information 2 .10 .12 .02 .18
Lonigan et al., 1991 Standard demographic information 1 .05
Nugent et al., 2007 Standard demographic information 2 .13 .08 .08 .19
Shannon et al., 1994 Standard demographic information 1 .06
Social withdrawal Stallard et al., 2001 Kidcope 2 .35 .01 .35 .36
Vernberg et al., 1996 Kidcope 1 .37
Thought suppression Aaron et al., 1999 White Bear Suppression Inventory 3 .62 .07 .55 .69

Stallard & Smith, 2007 Child rating of use of two strategies of
suppression
1 .59
Time Post-Trauma Goenjian et al., 2005 Time between event and assessment 3 −.29 .02 −.32 −.27
Koplewicz et al., 2002 Time between event and assessment 1 −.13
La Greca et al., 1996 Time between event and assessment 3 −.18 .08 −.27 −.12
Lengua et al., 2005 Time between event and assessment 1 −.06
Lengua et al., 2006 Time between event and assessment 1 −.03
Nugent et al., 2006 Time between event and assessment 1 −.15
Otto et al, 2007 Time between event and assessment 3 .15 .00 .15 .15
Schafer et al., 2006 Time between event and assessment 1 −.33
Thabet & Vostanis, 1999 Time between event and assessment 2 −.53 .06 −.57 −.49
Trauma Severity Aaron et al., 1999 Number of days in hospital 1 .32
Bal et al., 2003 Type of trauma 1 .30
Bradburn, 1991 Distance from epicenter 1 .57
Bryant et al, 2007b Injury severity 1 .10
Bryant et al., 2007 Injury severity 1 .91
Giannopoulou et al., 2006 Proximity to earthquake 1 .15
Goenjian et al., 1995 Proximity to epicenter 1 .50
Goenjian et al., 2001 Level of impact 3 .58 .16 .45 .76
Goenjian et al., 2005 Proximity to epicenter 6 .51 .16 .34 .71
Goenjian et al, 2009 Loss of parent 3 .09 .02 .07 .11
Heptinstall et al., 2004 Violent death of family member,
Experienced threats to life
Separated from parents
3 .24 .28 −.05 .50
Koplewicz et al., 2002 Directly witnessed event or not 4 .43 .16 .28 .64
La Greca et al., 1996 Number of life-threatening experiences 1 .39
Lack & Sullivan, 2008 Tornado Exposure Questionnaire 1 .11
Landolt et al., 2003 Days in hospital,
Functional status of child,
Type of trauma
3 .09 .15 −.08 .18

Landolt et al., 2005 Injury severity,
Days in hospital,
Parent involved in accident,
Type of accident
7 .02 .09 −.05 .21
Laor et al., 2002 Proximity to epicenter 1 .05
Linning & Kearney, 2004 Duration of maltreatment 1 .30
Lonigan et al., 1991 Hurricane Exposure 1 .23
Max et al., 1998 Glasgow Coma Scale 4 .24 .21 .05 .46
McDermott et al., 2005 Child endorsement of various trauma
variables
6 .26 .16 .11 .53
Mirza et al., 1998 Type of accident 1 .20
Nugent et al., 2007 Injury severity 4 .07 .04 .03 .11
Otto et al, 2007 Dose of exposure 2 .07 .05 .04 .11
Pfefferbaum et al., 1999 Relationship with deceased 2 .09 .01 .08 .10
Pfefferbaum et al., 2000 Indirect Interpersonal Exposure 1 .48
Pfefferbaum, 2001 Physical exposure 2 .18 .02 .17 .19
Rees et al., 2004 Severity of illness 1 .10
Roussos et al., 2005 Proximity to epicenter 1 .04
Runyon & Kenny, 2002 Trauma type 1 .40
Saxe et al., 2005 Surface of body area burned 2 .47 .03 .45 .49
Smith et al., 2001 Analogue Pain Scale 1 .28
Solomon & Lavi, 2005 Number of experienced traumatic events 3 .23 .02 .22 .25
Stallard & Smith, 2007 Type or RTA,
Parent involved in RTA,
Triage rating, A&E Outcome
4 .07 .06 .03 .13

A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents.pdf

A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents.pdf

Recommandé

Recommandé

Contenu connexe

Similaire à A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents.pdf

Similaire à A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents.pdf (20)

Plus de Tina Gabel

Plus de Tina Gabel (20)

Dernier

Dernier (20)

A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents.pdf